Self-assembly of nucleopeptides to interact with DNAs

Abstract

As a novel class of biomaterials, nucleopeptides, via the conjugation of nucleobases and peptides, usually self-assemble to form nanofibres driven mainly by hydrogen bonds. Containing nucleobase(s), nucleopeptides have a unique property—interacting with nucleic acids. Here we report the design and characterization of nucleopeptides that self-assemble in water and are able to interact with single-stranded DNAs (ssDNAs). Containing nucleobases on their side chains, these nucleopeptides bind with the ssDNAs, and the ssDNAs reciprocally affect the self-assembly of nucleopeptides. In addition, the interactions between nucleopeptides and ssDNAs also decrease their proteolytic resistance against proteinase K, which further demonstrates the binding with ssDNAs. The nucleopeptides also interact with plasmid DNA and deliver hairpin DNA into cells. This work illustrates a new and rational approach to create soft biomaterials by the integration of nucleobases and peptides to bind with DNA, which may lead to the development of nucleopeptides for controlling DNA in cells.

1. Introduction

Self-assembly, as a spontaneous process to organize molecules under thermodynamic equilibrium into well-defined structures and aggregated states through a number of non-covalent interactions. It occurs at all scales and is common throughout nature and has become a useful approach for developing nanotechnology [1–3]. Among these self-assembling molecules, peptide-based supramolecular hydrogelators, which consist of peptides [4–15] and a long alkyl chain [16–18], aromatic groups [19,20], or residues of opposite charge [21], can effectively generate hydrogels as soft biomaterials. Despite the successes of the use of peptides and peptidic derivatives for making hydrogels, there are still limitations (e.g. the length of the alkyl chain [18], the inhibition of cells [22,23] or the poor reversibility [24,25]). On the other hand, nucleobases, one of the fundamental building blocks chosen by nature for constructing nucleic acids, are capable of forming intermolecular interactions in water. Thus, nucleopeptides, which integrate peptides and nucleobases, are of considerable biological and biomedical importance [26]. For example, Algranati et al. reported that one kind of naturally occurring nucleopeptide, the spermidine-containing nucleopeptides or peptidyl nucleosides, can act as antibiotics against bacteria or fungi [27,28]. Buchardt and other groups reported that some unnatural nucleobase-containing peptides, such as polyamide nucleic acids (PNA), can mimic the structure and function of DNA with successful applications in biology and biomedicine [29–32]. Moreover, nucleopeptides are able to act as hydrogelators for generating supramolecular hydrogels [33,34].

Besides the biocompatible, biomedical and self-assembling properties (e.g. cell delivery [35], antibacterial [36], cancer cell inhibition [22] and scaffolds for regenerative medicine [37]), a unique ability of nucleopeptides is that their assemblies may bind with DNA to function like biomacromolecules. Some earlier works using the self-assembly of small, biologically important building blocks in water [38] for mimicking the structures and functions of biomacromolecules have underscored the feasibility and promise of this approach. For example, Shimizu et al. demonstrated that the self-assembly of nucleotide-appended bolaamphiphiles was capable of forming supramolecular helical nanofibres in the presence of specific target DNA (e.g. oligoadenylic acids) [39,40]. Barthelemy et al. reported the self-assembly of uridine phosphocholine amphiphiles to form diverse supramolecular structures (e.g. vesicles, fibres, hydrogels and organogels) [41,42]. Shinkai et al. described that a uracil-appended cholesterol gelator showed the excellent self-assembly ability in most organic solvents, which displays an increased stability of the molecular assembly by complementarily binding to oligonucleotide DNA via base-pairing [43–45].

Despite the existence of several well-characterized forms of nucleopeptides (chiral nucleopeptides, achiral nucleopseudopeptides, or peptidyl and amino nucleosides) [46], the exploration of nucleopeptides mainly focuses on the nucleopeptides that contain the nucleobases on the main chains of peptides. However, few works have yet explored the use of nucleobases for modifying the side chains of peptides [47,48]. It is necessary and beneficial to explore this new strategy to complement the existing ones for developing supramolecular hydrogels. Based on this principle, we have designed a novel nucleopeptide that has nucleobases on its side chains, and evaluated the self-assembly abilities of the nucleopeptide and studied the interaction between the assemblies of the nucleopeptide and DNA. Our studies reveal that the nucleopeptide is able to self-assemble in water to form nanofibres of around 6 nm in width and results in a hydrogel at a concentration of 1 wt%. Specifically, the nucleopeptide can bind with ssDNA to increase the rigidity of the supramolecular hydrogel, interact with plasmid DNA, and deliver hairpin DNA into cells. As the first report of nucleopeptide hydrogelators that contain nucleobases on the side chains, this work illustrates a facile and rational approach for creating nucleopeptides that act as a new class of supramolecular hydrogelators for developing sophisticated soft materials to control DNA behaviour in live cells.

2. Material and methods

2.1. Instruments

Samples were purified using a Water Delta600 HPLC system, equipped with an XTerra C18 RP column and an in-line diode array UV detector and dried with Labconco Freezone 4.5 Plus vacuum lyophilizer. Transmission electron microscopy (TEM) images were taken on Morgagni 268 transmission electron microscope. Rheological data were obtained using a TA ARES G2 rheometer with a 25 mm cone plate. Circular dichroism (CD) spectra were obtained using a Jasco J-810 spectropolarimeter. Dynamic light scattering (DLS) data were obtained using an ALV (Langen, Germany) goniometer and correlator system with a 22 mW HeNe (λ = 633 nm) laser and an avalanche photodiode detector. Fluorescent and optical images were taken on a Leica TCS SP2 Spectral Confocal Microscope. MTT assay for cell viability test was read using a DTX 880 multimode detector.

2.2. Materials

We designed and synthesized three different compounds for this study: 1, Nap-Phe-Phe-Lys(Thymine)-Gly-Lys(Cytosine)-Gly-Lys(Thymine)-OH (LC-MS: [M − 1]⁻ 1462.21, [M + 1]⁺ 1464.16; yield: 95%; electronic supplementary material, figure S1); 2, Nap-Phe-Phe-Lys-Gly-Lys-Gly-Lys-OH (LC-MS: [M − 1]⁻ 977.53, [M + 1]⁺ 979.35; yield: 98%; electronic supplementary material, figure S2); and 3, Phe-Phe-Lys(Thymine)-Gly-Lys(Cytosine)-Gly-Lys(Thymine)-Glucosamine (LC-MS: [M − 1]⁻ 1453.69, [M + 1]⁺ 1455.77; yield: 82%; electronic supplementary material, figure S3). All of the solvents and chemical reagents were used as received from the commercial sources without further purification unless otherwise noted. The ssDNAs were purchased from IDT (Integrated DNA Technologies). The sequence of DNA1 is: 5′-AGAAGAAGAAGA-3′ and the sequence of DNA2 is 5′-AAGAAAGAAAGAAAGAA-3′. Amino acids and HBTU were purchased form GL Biochem (Shanghai) Ltd. Thymine-1 acetic acid was purchased from Aldrich. Adenine was purchased from Acros Organics. Proteinase K was purchased from Sigma (more than 800 unit ml⁻¹). Other materials and solvents were purchased form Fisher Scientific.

2.3. Transmission electron microscopy

Solutions and supplies: 2.0% (w/v) uranyl acetate (UA) (prepare by dissolving 200 mg UA in 10 ml of ddH₂O); filter strips (prepared by cutting Whatman 1 filter paper into small slivers); grids (400 mesh copper grids coated with continuous thick carbon film approximately 35 nm in thickness, purchased from Pacific Grid Tech. Co.) [49]. Before placing sample solution/gels on the grid (3 µl, sufficient to cover the grid surface), we glow discharged the carbon-coated grids to increase their hydrophilicity. About 10 s later, we placed three large drops of ddH₂O on parafilm to let the grid touch the water drops with the sample-loaded surface facing the parafilm, and then tilted the grid and gently absorbed water from the edge of the grid using a filter paper. Immediately after rinsing, we placed three large drops of UA staining solution on parafilm to let the grid touch the staining solution drops with the sample-loaded surface facing the parafilm, and then tilted the grid and gently absorbed the stain solutions from the edge of the grid using a filter paper. After drying the grid in air and it was ready for EM imaging, we examined the grid as soon as possible.

2.4. Rheological measurements

Point five millilitres of the hydrogel samples were placed on the parallel plate. For the strain sweep test, the test range was from 0.1 to 11%, frequency equals 10 rad s⁻¹; and for frequency sweep test, the frequency range was from 0 to 110 rad s⁻¹. Sweep mode was ‘log’ and temperature was carried out at 25°C.

2.5. Circular dichroism spectra

CD spectra (185–500 nm) were recorded by using a JASCO 810 spectrometer under a nitrogen atmosphere. The samples (0.2 ml, 1 wt %) were placed evenly on the 1 mm thick quartz curvet and scanned with 0.5 nm interval. We got the average spectra after three tests.

2.6. Dynamic light scattering

1 at a concentration of 500 µM was filtered by using 0.22 µm filters after heating. The incubation of ssDNAs (i.e. DNA1 or DNA2) with the solution of 1 at room temperature for 24 h generated the corresponding samples. The DLS tests were carried out at room temperature, and the angle of light scattering we chose was 90°. The resulting intensity ratios were proportional to the amount of aggregates in the samples.

2.7. Biostability test with proteinase K

One milligram of 1 or 1 with DNA1 or DNA2 was dissolved in 5 ml HEPES buffer at pH = 7.5. Then proteinase K was added at a concentration of 3.2 units ml⁻¹ and incubated at 37°C for 24 h, then 100 µl of the sample was taken out each time and analysed using HPLC.

2.8. DNA gel electrophoresis

After mixing 1 g of agarose powder with 100 ml 1 × TAE buffer in a microwavable flask, we microwaved the mixture for 3 min until the agarose was completely dissolved. After the agarose solution cooling down to about 50°C, we added ethidium bromide to a final concentration of approximately 0.2 µg ml⁻¹. We poured the agarose into a gel tray with the well comb in place and placed the newly poured gel at 4°C for 10–15 min to make it completely solidified. We prepared the samples and then added the loading buffer. Once the gel was solidified, we placed the agarose gel into the gel box and filled the gel box with 1 × TAE buffer until the gel was covered. After carefully loading the ladder and the samples into the wells of the gel, we ran the gel at 100 V for 1.5 h until the dye line was approximately 75–80% of the way down the gel. After turning off power, we disconnected the electrodes from the power source, and then carefully removed the gel from the gel box. The gel was ready for film.

2.9. Cells culture

HeLa cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). HeLa cells were propagated in minimum essential media (MEM, Invitrogen Life Technologies) supplemented with 10% fetal bovine serum (FBS, Invitrogen Life Technologies) and 1% of antibiotics (100 U ml⁻¹ penicillin and 100 µg ml⁻¹ streptomycin, Invitrogen Life Technologies). The cells were incubated in a fully humidified incubator containing 5% CO₂ at 37°C.

2.10. Live cell imaging

HeLa cells in exponential growth phase were seeded in glass bottomed culture chambers at 1 × 10⁵ cells well⁻¹. The cells were allowed to attach for 24 h at 37°C, 5% CO₂. The culture medium was removed, and new culture medium containing 1 or 2 at 15.7 µM and hairpin DNA at 250 nM was added. After incubation for 1 h, HeLa cells were rinsed three times with PBS buffer, and then kept in PBS buffer for imaging.

2.11. Cell viability assay

Cells in exponential growth phase were seeded in a 96-well plate at a density of 1 × 10⁴ cell well⁻¹. The cells were allowed to attach to the wells for 4 h at 37°C, 5% CO₂. The culture medium was removed and 100 µl culture medium containing the compounds (immediately diluted from fresh prepared stock solution of 10 mM) at gradient concentrations (0 µM as the control) was placed into each well. After culturing at 37°C, 5% CO₂ for 24, 48, 72 h, 10 µl of 5 mg ml⁻¹ MTT (3-(4,5-DimethylthiazoL-2-yl)-2,5-diphenyltetrazolium bromide) was added to each well, and the plated cells were incubated in the dark for 4 h. One hundred microlitres of 10% SDS with 0.01 M HCl was added to each well to stop the reduction reaction and to dissolve the purple. After incubation of the cells at 37°C for overnight, the OD at 595 nm of the solution was measured in a microplate reader (DTX 880 multimode detector). Data represented the mean ± standard deviation of three independent experiments.

3. Results and discussion

Scheme 1 shows the structure of the nucleopeptide (1, Nap-Phe-Phe-Lys(Thymine)-Gly-Lys(Cytosine)-Gly-Lys(Thymine)-OH), consisting of N-terminal capped peptide and three nucleobases (two thymines and one cytosine) on its side chain. The introduction of nucleobases at the side chains of 1 allows it to bind with single stranded DNAs (ssDNAs), which would affect the self-assembly of 1 to form the nanostructures of different morphologies (from the single nanofibre to the clustering of nanofibres) and increase the rigidity of hydrogels. The existence of the naphthalene (Nap) group increases the self-assembly ability of 1 through π–π interactions, which contributes to the binding between nucleopeptides and DNAs. As the control, we designed another peptide (2, Nap-Phe-Phe-Lys-Gly-Lys-Gly-Lys-OH, electronic supplementary material, scheme S1) and examined the interactions between 2 and ssDNAs.

Scheme 1.

Illustration of the structure and self-assembly of 1 and the binding between 1 and ssDNAs.

As shown in scheme 2, the synthesis of the designed nucleopeptide 1 is fast and straightforward. We first use N-hydroxysuccinimide (NHS) to activate the carboxylic acid groups on the nucleobases for directly coupling them with the ε-amine on the side chain of a lysine [50]. Then, by using solid-phase peptide synthesis (SPPS), we obtain the nucleopeptide, 1, directly. After the synthesis and purification of nucleopeptide 1, we investigated the self-assembly ability of 1 and the interaction between 1 and various DNAs (e.g. ssDNAs, plasmid DNA or hairpin DNA).

Scheme 2.

The chemical synthesis route of 1.

We used TEM to evaluate the morphological change of nanostructures formed by the gel with 1 before or after the addition of ssDNAs (figure 1). We designed two different ssDNA sequences (i.e. DNA1 or DNA2) with the different adenine–guanine–adenine (AGA) spacing, which should provide two kinds of binding models. Since we have little a priori knowledge about the positions of thymine–cytosine–thymine (TCT) after the self-assembly of 1, we involve two ssDNA sequences, which consist of the AGA domains with different spacing, to bind with various TCT. The interbase spacing of AGA in DNA1 is smaller than that of DNA2, while the latter exhibits the similar interbase spacing of the TCT sequence on the side chain of 1. Thus, we suppose that DNA2 would show better binding affinity with the assemblies of 1. As shown in figure 1, self-assembly of 1 (1 wt%) can form a weak (i.e. semi-fluid) gel at pH 7.4 consisting of uniform nanofibres with the diameter of 6 ± 2 nm. The addition of ssDNAs contributes to forming strong hydrogels at the same concentration and inducing the clustering of several nanofibres (figure 1b,c). The binding between ssDNAs and the nucleobases on the nucleopeptides increases the rigidity of the hydrogels and clusters the single nanofibre together by ssDNAs. In addition, there are more aggregates (likely to be the ssDNAs themselves, electronic supplementary material, figure S4) in the hydrogel of 1 and DNA1 than in that of 1 and DNA2. The interbase spacing of AGA in DNA2 (5′-AAGAAAGAAAGAAAGAA-3′) is comparable with that of the TCT sequence on the side chain of 1, which results in the enhanced binding between 1 and DNA2. However, the interbase spacing of DNA1 (5′-AGAAGAAGAAGA-3′) is smaller than that of the TCT sequence. This fact, together with the TEM images of the mixture hydrogels, further suggests that 1 interacts more strongly with DNA2 than with DNA1.

Figure 1.

Transmission electron microscopy (TEM) images of the gel formed by (a) 1; (b) 1 with DNA1; (c) 1 with DNA2 at a concentration of 1.0 wt%, and pH = 7.0. Inserted are the optical images of their corresponding hydrogels. The concentration of DNAs is 1.75 mM. The scale bar is 100 nm.

In order to investigate the viscoelastic properties, one of the essential features of hydrogels, we performed dynamic rheological experiments to study the viscoelasticity of the hydrogels consisting of 1 or 2 and ssDNAs. The storage modulus (G′) and loss modulus (G″) indicate the ability of the deformed hydrogel to restore its original geometry and the tendency of a material to flow under exerted stress, among which G′ is usually greater than G″ for a gel system. Based on the crossover between G′ and G″, we can determine the critical gelation concentration (cgc) of the samples. The cgc of 1 is 0.07 wt% (electronic supplementary material, figure S5). However, the addition of ssDNAs will slightly increase the cgc of 1 (from 0.07 wt% to 0.09 wt%, electronic supplementary material, figure S6), indicating that the interactions between 1 and ssDNAs may destabilize the self-assembly of 1 at low concentrations. Since we prepare the hydrogels or hydrogels with ssDNAs under physiological pHs (i.e. pH = 7.4), the formed hydrogels likely would be stable at concentrations higher than 0.1 wt% for the studies on the binding with DNAs or potential gene delivery. In addition, the cgc of 2 is 0.6 wt% (electronic supplementary material, figure S7). Because of the positive charges in 2, the addition of ssDNAs results in the precipitates and the bundles of nanofibres (electronic supplementary material, figure S8).

In the oscillatory strain sweep experiment, we note that the hydrogel of 1 and DNA2 possesses the highest maximum storage modulus of around 1.8 × 10³ Pa (figure 2a) at a concentration of 1 wt%. The storage modulus of hydrogel of 1 and DNA1 (1.1 × 10³ Pa) is slightly lower than that of 1 and DNA2, while the storage modulus of hydrogel of 1 is the lowest (0.5 × 10³ Pa). In addition, the critical strains of these hydrogels gradually increase and are 5.21, 6.34, and 7.03%, respectively, from hydrogel of 1 to 1 and DNA2, suggesting increasing steric stability. These results indicate that the hydrogel of 1 and DNA2 has the most solid-like rheological behaviour and is consistent with the TEM results. Frequency sweep (figure 2b) shows that the storage modulus is always higher than the loss modulus for each hydrogel, which is consistent with the viscoelasticity of the hydrogels. However, for the control peptide 2, because of the formation of precipitates, the rigidity of 2 slightly decreases with the addition of ssDNAs (electronic supplementary material, figure S9).

Figure 2.

(a) Strain and (b) frequency dependence of the dynamic storage moduli (G′) and the loss moduli (G″) the gel of 1 at the concentration of 1.0 wt%, the gel of 1 with DNA1 and that with DNA2.

To further evaluate the interaction between 1 and ssDNAs, we examined the mixture of hydrogels using circular dichroism (CD) (figure 3). The peak at 310 nm results from the three nucleobases on the side chain of 1 [51]. The CD spectra of 1 displays a large negative band at 310 nm while the addition of DNA1 in the gel of 1 induces a smaller negative band and the addition of DNA2 results in a positive band (figure 3a), which indicates the interaction between the nucleobases of 1 and ssDNAs. The binding between 1 and ssDNAs decreases the amount of free TCT sequences. As a result, it changes the conformation of the mixture to switch the negative band at 310 nm to a positive band. In addition, based on the CD spectra from 190 nm to 240 nm (inserted in figure 3a), after adding DNA1 or DNA2 into the hydrogel of 1, the structure changes from a β-sheet to a α-helix-like structure or to a random coil, likely due to the helical conformation of ssDNAs. These results further demonstrate the binding between 1 and ssDNAs.

Figure 3.

(a) Circular dichroism (CD) spectra for the gel of 1 at a concentration of 1.0 wt%, the gel of 1 with DNA1 and the gel of 1 with DNA2. Inserted is the enlarged graph of spectra from 185 nm to 240 nm. (b) The intensity of dynamic light scattering (DLS) of 1 at a concentration of 500 µM, 1 with DNA1, and 1 with DNA2 (the ratio of 1 and DNAs is 1 : 1) at an angle of 90°. Data are shown as mean ± s.d. **p < 0.01, ***p < 0.001 by Student's t test. n = 3.

Besides studying the secondary structures of the mixture hydrogels, we used DLS, which measures the correlation of scattered light for detecting the concentration fluctuation induced by particles that diffuse in solution, to investigate the effect of ssDNAs on the self-assembly ability of 1 at concentrations below the gelation concentration. Thus, we tested the DLS intensity at a relatively lower concentration of 1 (i.e. 500 µM). As shown in figure 3b, the addition of DNA1 to the solution of 1 only increases the DLS intensity twofold while the addition of DNA2 increases the intensity around ninefold. This result confirms that the binding between 1 and DNA2 is stronger than that of 1 and DNA1, which is further demonstrated by the next morphological studies on the mixtures of ssDNAs and 1 at a concentration of 500 µM.

Interestingly, in the study of the interaction between 1, at a concentration of 500 µM, and these two ssDNAs, we find that the addition of DNA2 generates many short nanofibres, while the solution of 1 alone contains long nanofibres of uniform width (figure 4). The solution of 1 and DNA1 consists of similar nanofibres to the solution of 1 with a width of around 6 nm, and contains some small aggregates. However, the addition of DNA2 significantly changes the morphology of nanofibre from several microns to several hundred nanometres in length. These results explain the dramatically increased DLS intensity of 1 and DNA2 (figure 3), in comparison with 1 itself. In addition, the end of one short nanofibre formed by 1 and DNA2 usually connects with another nanofibre (figure 4c), which likely resulted from the binding of nucleobases and ssDNAs.

Figure 4.

TEM images of (a) 1 at a concentration of 500 µM; (b) 1 with DNA1; (c) 1 with DNA2 (the ratio of 1 and DNAs is 1 : 1). The scale bar is 100 nm.

Supramolecular hydrogels, made from small peptides or short l-amino acid sequences, are usually susceptible to biodegradation through hydrolysis catalysed by proteolytic enzymes [52,53]. We test the biostability of 1 and the mixture of 1 and ssDNAs by incubating the hydrogelators with proteinase K (figure 5), a powerful protease that degrades a wide range of peptidic substrates [54,55]. As shown in figure 5, we use 3.2 unit ml⁻¹ proteinase K to treat 1 dissolved in HEPES buffer at a concentration of 0.02 wt%. We find that around 50% of 1 remains after 1 h incubation with proteinase K, and 40% of 1 is left with 8 h incubation. However, the addition of ssDNAs (i.e. DNA1 or DNA2) causes 1 to be complete digested in 1 h whether the ratio of 1 to ssDNAs is 1 : 1 (figure 5) or 10 : 1 (electronic supplementary material, figure S10). The interaction between 1 and ssDNAs apparently destabilizes the nanofibres of 1 at lower concentrations and accelerates the digestion of 1 by proteinase K, which is similar to the degradation of peptide nanofibrils via supramolecular glycosylation [56]. At a concentration of 0.02 wt%, the binding of 1 and ssDNAs makes 1 more easily released from the formed nanostructures and degraded quicker by proteinase K, which agrees with the TEM images in figure 4 that DNA2 breaks and shortens the nanofibres of 1 at a concentration of 500 µM (0.073 wt%).

Figure 5.

Digestion curve of 1; 1 with DNA1; 1 with DNA2 (the ratio of 1 and DNAs is 1 : 1) upon the treatment of proteinase K. The concentration of proteinase K is 3.2 unit ml⁻¹. Data are shown as mean ± s.d. *p < 0.05, **p < 0.01 by Student's t test. n = 3.

Our studies show that the interactions with ssDNAs are able to increase the self-assembly ability of 1 and to enhance the rigidity of the hydrogels while apparently destabilizing the nanofibres of 1 at lower concentrations. Besides focusing on the interaction with ssDNAs, we also want to expand this property to other more complex DNAs. Thus, we study the interactions between 1 and other DNAs (e.g. plasmid DNA or hairpin DNA). The DNA gel electrophoresis analysis shows that the band of the free plasmid DNA becomes weaker with the increase of the concentration of 1 from 0.25 mM to 10 mM (figure 6a). Together with the result that more plasmid DNA is trapped at the starting point, we can conclude that 1 is able to bind with plasmid DNA, which will be proportional to the concentration of 1. In addition, the incubation time of 1 with plasmid DNA seems to have little effect on the binding since 30 min or 1 s incubation shows similar plasmid DNA bands. The strain sweep and frequency sweep results indicates that adding plasmid DNA into the gel of 1 increases the storage modulus fourfold (figure 6b,c), agreeing with the interaction between 1 and plasmid DNA. This phenomenon results from the binding between 1 and plasmid DNA and the stronger self-assembly abilities of the mixture. Since we prepare the samples at pH 7.4, the nucleopeptides should contain the negative charge. The increased plasmid DNA trapped at the starting point unlikely originates from the isoelectric point of the mixture.

Figure 6.

(a) DNA gel electrophoresis analysis of the interaction between 1 and plasmid DNA. Lanes 1 to 10 contain the ladder (1), 0.25 mM, 0.5 mM, 1 mM, 2.5 mM, 5 mM, or 10 mM 1 + plasmid DNA (3.9 µg) (2–7, all samples are incubated for 30 min), just plasmid DNA (8), 10 mM 1 + plasmid DNA (3.9 µg) incubated for 1 s (9), and 10 mM 1 alone (10). (b) Strain and (c) frequency dependence of the dynamic storage moduli (G′) and the loss moduli (G″) the gel of 1 at a concentration of 1.0 wt%, the gel of 1 with plasmid DNA.

To further explore the interaction between 1 and DNAs, an attractive feature of this kind of hydrogelators, we investigated whether 1 facilitates the delivery of DNAs into live cells and examined the subcellular distribution of the delivered DNAs (figure 7). We designed another nucleopeptide (3, electronic supplementary material, scheme S1) by connecting one sugar with the C-terminus of the peptide chain to increase its ability to deliver DNAs into live cells. We incubated 1 or 3 at a concentration of 15.7 µM with 250 nM of hairpin DNA labelled with a fluorescent dye (Cy5-labelled hairpin DNA). After 1 h of the incubation, we removed the culture medium, washed the cells with PBS buffer, and took fluorescent images. As shown in figure 7, with the assistance of 1 or 3, the red fluorescence shows in the cytosols of HeLa cells, indicating the presence of hairpin DNA in the HeLa cells. In the control experiment (i.e. without using nucleopeptides), red fluorescence is absent from the cytosols of the HeLa cells, indicating that it is the nucleopeptide that interacts with and delivers hairpin DNA into the live cells. Based on the fluorescent images, 3 shows a better ability to deliver hairpin DNA into the live cells than 1 because of its better solubility. In addition, we verified the biocompatibility of the nucleopeptides by adding 1 or 3 into a culture of mammalian cells and measuring the proliferation of the cells. According to the results of the MTT assay (electronic supplementary material, figure S11), even after being incubated with 500 μM of the nucleopeptides for 72 h, the cell viability remains at around 100%, indicating that 1 and 3 are biocompatible.

Figure 7.

HeLa cells were treated with 1 and hairpin DNA, 3 and hairpin DNA, or hairpin DNA itself (control) for 1 h. The concentration of hairpin DNA was 250 nM, and the concentration of the compounds was 15.7 µM. The scale bar is 50 µm.

4. Conclusion

In conclusion, we have developed a new type of nucleopeptide by conjugating nucleobases to peptide derivative side chains. We have demonstrated that the designed nucleopeptides can bind with ssDNAs to enhance its self-assembly ability, interact with plasmid DNA, and help deliver hairpin DNA into live cells. In addition, the biocompatibilities of these nucleopeptides promise more potential functions, such as binding with double stranded DNAs, DNA origami, or interacting with DNA in live cells. This work not only introduce a facile way to expand the current repertoire of nucleopeptides for generating supramolecular assemblies from biomolecules but also provides a possible approach for designing soft biomaterials to manipulate DNA behaviour and function in live cells.

Data accessibility

This article has no additional data.

Authors' contributions

B.X. conceptually designed the strategy for this study, provided intellectual input, supervised the studies and wrote the manuscript. X.D. designed and performed experiments of chemical synthesis, in vitro operation, analysed the experimental results, generated the figures and wrote the manuscript. J.Z. helped design the experiments and drew the scheme. X.L. synthesized some compounds.

Competing interests

The authors declare no competing financial interests.

Funding

This work is partially supported by NIH (R01CA142746) and NSF (DMR-1420382). J.Z. is an HHMI International Research Fellow.